Psychological constructs are reliable using one of

This preview shows page 14 - 16 out of 28 pages.

We have textbook solutions for you!
The document you are viewing contains questions related to this textbook.
Introduction to Probability and Statistics
The document you are viewing contains questions related to this textbook.
Chapter 5 / Exercise 22
Introduction to Probability and Statistics
Beaver/Mendenhall
Expert Verified
psychological constructs are reliable using one of several types of tests. Test-Retest Reliability. The objective with test-retest reli- ability assessment is to measure the same subjects at two dif- ferent points in time under approximately the same conditions and compare results. If there was actually no change in the object of measurement (such as someone’s preferences or at- titudes), a reliable measuring device (such as a questionnaire) would generate the same results. Obvious problems are that it may not be possible to get the same respondents to respond to the test, the measurement process itself may alter the second responses, and conditions may be difficult to duplicate. Alternative Form. The deficiencies of test-retest reliabil- ity measures can be overcome to some degree by using an alternative form of the measuring device. If you have at least two equivalent measuring forms (such as questionnaires), re- searchers can administer one form and then two weeks later administer the alternative form. Correlating the results provides a measure of the reliabil- ity of the forms. Internal Consistency. Measuring reliability using an internal consistency approach usually involves the use of different samples of respondents to determine reliability of measures. A typical way to assess internal consistency would be to use a split-half tech- nique, which involves giving the measuring instrument (such as a questionnaire) to all the respondents, then randomly dividing them into halves and comparing the two halves. High correlation represents good reliability. Another type of split-half comparison in- volves randomly splitting the item statements intended to measure the same construct into halves and comparing the results of the two halves of the statements for the entire sample. Again, high correlation of results (the way the respondents answered the item statements) suggests good reliability. Another way of testing for internal consistency is to use a statistical test such as Cronbach’s alpha for intervally scaled data or KR-20 for nom- inally scaled data. These measures will indicate how well each individual item statement correlates with other items in the instrument. A low correlation means the item statement visivastudio/Shutterstock.com
We have textbook solutions for you!
The document you are viewing contains questions related to this textbook.
Introduction to Probability and Statistics
The document you are viewing contains questions related to this textbook.
Chapter 5 / Exercise 22
Introduction to Probability and Statistics
Beaver/Mendenhall
Expert Verified
Chapter 6 Measurement 111 should be removed from the instrument. Cronbach’s alpha is a common test done in Step 5: Purify the Measures of Our Measurement Process (see Figure 6.2). Also note that in Figure 6.1, reliability tests are using measures (data such as answers to questionnaires) to test whether our operational definitions (questionnaire questions) are reliable (e.g., internally consistent). Validity Validity is the extent to which differences found among respondents using a measuring tool reflect true differences among respondents. The difficulty in assessing validity is that the true value is usually unknown. If the true value were known, absolute validity could be measured. In the absence of knowledge of the true value, the concern must be with relative validity, that is, how well the variable is measured using one measuring technique

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture